Read a scan report

Img DeDup can export what a scan saw. Drop those files here and this page turns them into a picture of your library — how much of it is genuinely duplicated, how much the scan could not see at all, and what is worth deleting first.

Upload your diagnostic files

Everything the app can export is accepted. They are read here in your browser to build the report — none of them is sent anywhere.

diagnostics.json Required One row per item from the last scan. This is what the duplicate tiers, the chart and the advice are computed from.
snapshots.json Optional Library size at each recorded scan. Adds a chart of your storage over time.
deletion-log.json Optional Copies already deleted, and the original kept in each case. Adds an "already reclaimed" total.
Savings report Optional The plain-text summary from the Savings screen. Reproduced at the end of the report.

Drop your export here

Or choose files. You can select more than one at a time.

Nothing is uploaded

Your file is read by this page, in this tab, and analysed here. It is not sent anywhere — and there is nowhere for it to be sent: this site is served as static files by a Worker that has no server code and no storage attached to it. Close the tab and nothing of yours remains.

Where the files come from

In the app, open Scan and look under Diagnostics for Share diagnostics.json. The savings report, deletion-log.json and snapshots.json are on the Savings screen. On a phone you can also tap Analyse on the web and skip the sharing entirely.

The same arithmetic

The grouping here is the app's own, not a rough version of it: the same fingerprint, the same distance thresholds, the same rule about which copy is kept. The numbers on this page and the numbers on your phone should agree.

How to read it

The report is deliberately arranged as a ladder rather than a single number, because the interesting question is not how many duplicates exist. It is where a reasonable person stops.

Exact copies: Two files whose original bytes hash to the same value. These are the same file twice, and nothing is lost by keeping one. There is no judgement to make here.
Identical fingerprints: Different bytes, same picture, down to the last compared pixel of the reduction. This is what a re-save, a re-download or an export produces. Safe in practice.
Strict and normal matches: Fingerprints that differ by a couple of bits: a crop, a filter, a resize, a recompression. Usually the same photo, and worth looking at before deleting — which is what the app's set view is for.
Loose matches: The widest net. It catches variants the other tiers miss and it also catches genuinely different pictures that happen to share a composition — five frames of the same landscape, say. Review this one properly or leave it alone.
What the scan could not see: An asset with no local preview gets no fingerprint, and video is never fingerprinted at all. If that is most of your storage, duplicates are not your problem and this page will say so rather than quietly leaving it out of the total.